diff --git a/swh/storage/archiver/director.py b/swh/storage/archiver/director.py index a9f630f7..e0101e0a 100644 --- a/swh/storage/archiver/director.py +++ b/swh/storage/archiver/director.py @@ -1,287 +1,296 @@ # Copyright (C) 2015-2016 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import logging import click import time from swh.core import hashutil, config from swh.objstorage import PathSlicingObjStorage from swh.objstorage.api.client import RemoteObjStorage from swh.scheduler.celery_backend.config import app from . import tasks # NOQA from .storage import ArchiverStorage DEFAULT_CONFIG = { 'objstorage_type': ('str', 'local_storage'), 'objstorage_path': ('str', '/tmp/swh-storage/objects'), 'objstorage_slicing': ('str', '0:2/2:4/4:6'), 'objstorage_url': ('str', 'http://localhost:5003/'), 'batch_max_size': ('int', 50), 'archival_max_age': ('int', 3600), 'retention_policy': ('int', 2), 'asynchronous': ('bool', True), 'dbconn': ('str', 'dbname=softwareheritage-archiver-dev user=guest') } task_name = 'swh.storage.archiver.tasks.SWHArchiverTask' logger = logging.getLogger() class ArchiverDirector(): """Process the files in order to know which one is needed as backup. The archiver director processes the files in the local storage in order to know which one needs archival and it delegates this task to archiver workers. Attributes: master_objstorage: the local storage of the master server. master_objstorage_args (dict): arguments of the master objstorage initialization. archiver_storage: a wrapper for archiver db operations. db_conn_archiver: Either a libpq connection string, or a psycopg2 connection for the archiver db. slave_objstorages: Iterable of remote obj storages to the slaves servers used for backup. config: Archiver_configuration. A dictionary that must contain the following keys: objstorage_type (str): type of objstorage used (local_storage or remote_storage). If the storage is local, the arguments keys must be present objstorage_path (str): master's objstorage path objstorage_slicing (str): masters's objstorage slicing Otherwise, if it's a remote objstorage, the keys must be: objstorage_url (str): url of the remote objstorage batch_max_size (int): The number of content items that can be given to the same archiver worker. archival_max_age (int): Delay given to the worker to copy all the files in a given batch. retention_policy (int): Required number of copies for the content to be considered safe. asynchronous (boolean): Indicate whenever the archival should run in asynchronous mode or not. """ def __init__(self, db_conn_archiver, config): """ Constructor of the archiver director. Args: db_conn_archiver: Either a libpq connection string, or a psycopg2 connection for the archiver db. config: Archiver_configuration. A dictionary that must contain the following keys: objstorage_type (str): type of objstorage used (local_objstorage or remote_objstorage). If the storage is local, the arguments keys must be present objstorage_path (str): master's objstorage path objstorage_slicing (str): masters's objstorage slicing Otherwise, if it's a remote objstorage, the keys must be: objstorage_url (str): url of the remote objstorage batch_max_size (int): The number of content items that can be given to the same archiver worker. archival_max_age (int): Delay given to the worker to copy all the files in a given batch. retention_policy (int): Required number of copies for the content to be considered safe. asynchronous (boolean): Indicate whenever the archival should run in asynchronous mode or not. """ # Get the slave storages self.db_conn_archiver = db_conn_archiver self.archiver_storage = ArchiverStorage(db_conn_archiver) self.slave_objstorages = { id: url for id, url in self.archiver_storage.archive_ls() } # Check that there is enough backup servers for the retention policy if config['retention_policy'] > len(self.slave_objstorages) + 1: raise ValueError( "Can't have a retention policy of %d with %d backup servers" % (config['retention_policy'], len(self.slave_objstorages)) ) # Get the master storage that contains content to be archived if config['objstorage_type'] == 'local_objstorage': master_objstorage_args = { 'root': config['objstorage_path'], 'slicing': config['objstorage_slicing'] } master_objstorage = PathSlicingObjStorage( **master_objstorage_args ) elif config['objstorage_type'] == 'remote_objstorage': master_objstorage_args = {'base_url': config['objstorage_url']} master_objstorage = RemoteObjStorage(**master_objstorage_args) else: raise ValueError( 'Unknow objstorage class `%s`' % config['objstorage_type'] ) self.master_objstorage = master_objstorage self.master_objstorage_args = master_objstorage_args # Keep the full configuration self.config = config def run(self): """ Run the archiver director. The archiver director will check all the contents of the archiver database and do the required backup jobs. """ if self.config['asynchronous']: run_fn = self.run_async_worker else: run_fn = self.run_sync_worker for batch in self.get_unarchived_content(): run_fn(batch) def run_async_worker(self, batch): """ Produce a worker that will be added to the task queue. """ task = app.tasks[task_name] task.delay(batch, archiver_args=self.db_conn_archiver, master_objstorage_args=self.master_objstorage_args, slave_objstorages=self.slave_objstorages, config=self.config) def run_sync_worker(self, batch): """ Run synchronously a worker on the given batch. """ task = app.tasks[task_name] task(batch, archiver_args=self.db_conn_archiver, master_objstorage_args=self.master_objstorage_args, slave_objstorages=self.slave_objstorages, config=self.config) def get_unarchived_content(self): """ Get contents that need to be archived. Yields: A batch of contents. Batches are dictionaries which associates a content id to the data about servers that contains it or not. {'id1': {'present': [('slave1', 'slave1_url')], 'missing': [('slave2', 'slave2_url'), ('slave3', 'slave3_url')] }, 'id2': {'present': [], 'missing': [ ('slave1', 'slave1_url'), ('slave2', 'slave2_url'), ('slave3', 'slave3_url') ]} } Where keys (idX) are sha1 of the content and (slaveX, slaveX_url) are ids and urls of the storage slaves. At least all the content that don't have enough copies on the backups servers are distributed into these batches. """ contents = {} # Get the archives - archives = [(k, v) for k, v in self.archiver_storage.archive_ls()] + archives = dict(self.archiver_storage.archive_ls()) # Get all the contents referenced into the archiver tables - for content_id, copies in self.archiver_storage.content_archive_get(): - content_id = r'\x' + hashutil.hash_to_hex(content_id) - data = {'present': [], 'missing': []} - # For each archive server, check the current content status - for archive_id, archive_url in archives: - if archive_id not in copies: - data['missing'].append((archive_id, archive_url)) - else: - copy = copies[archive_id] - vstatus = self.get_virtual_status(copy['status'], - copy['mtime']) - data[vstatus].append((archive_id, archive_url)) - - contents[content_id] = data - - if len(contents) >= self.config['batch_max_size']: - yield contents - contents = {} + last_object = b'' + while True: + archived_contents = list( + self.archiver_storage.content_archive_get_copies(last_object)) + + if not archived_contents: + break + + for content_id, present, ongoing in archived_contents: + last_object = content_id + data = { + 'present': set(present), + 'missing': set(archives) - set(present) - set(ongoing), + } + + for archive_id, mtime in ongoing.items(): + status = self.get_virtual_status('ongoing', mtime) + data[status].add(archive_id) + + contents[r'\x%s' % hashutil.hash_to_hex(content_id)] = { + k: [(archive_id, archives[archive_id]) for archive_id in v] + for k, v in data.items() + } + + if len(contents) >= self.config['batch_max_size']: + yield contents + contents = {} if len(contents) > 0: yield contents def get_virtual_status(self, status, mtime): """ Compute the virtual presence of a content. If the status is ongoing but the time is not elasped, the archiver consider it will be present in the futur, and so consider it as present. However, if the time is elasped, the copy may have failed, so consider the content as missing. Arguments: status (string): One of ('present', 'missing', 'ongoing'). The status of the content. mtime (datetime): Time at which the content have been updated for the last time. Returns: The virtual status of the studied content, which is 'present' or 'missing'. Raises: ValueError: if the status is not one 'present', 'missing' or 'ongoing' """ if status in ('present', 'missing'): return status # If the status is 'ongoing' but there is still time, another worker # may still be on the task. if status == 'ongoing': elapsed = int(time.time()) - mtime if elapsed <= self.config['archival_max_age']: return 'present' else: return 'missing' else: raise ValueError("status must be either 'present', 'missing' " "or 'ongoing'") @click.command() @click.argument('config-path', required=1) @click.option('--dbconn', default=DEFAULT_CONFIG['dbconn'][1], help="Connection string for the archiver database") @click.option('--async/--sync', default=DEFAULT_CONFIG['asynchronous'][1], help="Indicates if the archiver should run asynchronously") def launch(config_path, dbconn, async): # The configuration have following priority : # command line > file config > default config cl_config = { 'dbconn': dbconn, 'asynchronous': async } conf = config.read(config_path, DEFAULT_CONFIG) conf.update(cl_config) # Create connection data and run the archiver. archiver = ArchiverDirector(conf['dbconn'], conf) logger.info("Starting an archival at", time.time()) archiver.run() if __name__ == '__main__': launch() diff --git a/swh/storage/archiver/storage.py b/swh/storage/archiver/storage.py index eb5f031c..9a62f67a 100644 --- a/swh/storage/archiver/storage.py +++ b/swh/storage/archiver/storage.py @@ -1,73 +1,93 @@ # Copyright (C) 2016 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import psycopg2 from ..common import db_transaction_generator, db_transaction from ..db import Db from ..exc import StorageDBError class ArchiverStorage(): """SWH Archiver storage proxy, encompassing DB """ def __init__(self, db_conn): """ Args: db_conn: either a libpq connection string, or a psycopg2 connection """ try: if isinstance(db_conn, psycopg2.extensions.connection): self.db = Db(db_conn) else: self.db = Db.connect(db_conn) except psycopg2.OperationalError as e: raise StorageDBError(e) @db_transaction_generator def archive_ls(self, cur=None): """ Get all the archives registered on the server. Yields: a tuple (server_id, server_url) for each archive server. """ yield from self.db.archive_ls(cur) @db_transaction_generator def content_archive_get(self, content=None, cur=None): """ Get the archival status of a content in a specific server. Retreive from the database the archival status of the given content in the given archive server. Args: content: the sha1 of the content. May be None for any id. archive: the database id of the server we're looking into may be None for any server. Yields: A tuple (content_id, server_id, archival status, mtime, tzinfo). """ yield from self.db.content_archive_get(content, cur) + @db_transaction_generator + def content_archive_get_copies(self, previous_content=None, limit=1000, + cur=None): + """Get the list of copies for `limit` contents starting after + `previous_content`. + + Args: + previous_content: sha1 of the last content retrieved. May be None + to start at the beginning. + limit: number of contents to retrieve. Can be None to retrieve all + objects (will be slow). + + Yields: + A tuple (content_id, present_copies, ongoing_copies), where + ongoing_copies is a dict mapping copy to mtime. + + """ + yield from self.db.content_archive_get_copies(previous_content, limit, + cur) + @db_transaction def content_archive_update(self, content_id, archive_id, new_status=None, cur=None): """ Update the status of an archive content and set its mtime to Change the mtime of an archived content for the given archive and set it's mtime to the current time. Args: content_id (str): content sha1 archive_id (str): name of the archive new_status (str): one of 'missing', 'present' or 'ongoing'. this status will replace the previous one. If not given, the function only change the mtime of the content for the given archive. """ self.db.content_archive_update(content_id, archive_id, new_status, cur) diff --git a/swh/storage/db.py b/swh/storage/db.py index 03614346..e2cb01ce 100644 --- a/swh/storage/db.py +++ b/swh/storage/db.py @@ -1,719 +1,769 @@ # Copyright (C) 2015 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import binascii import datetime import functools import json import psycopg2 import psycopg2.extras import tempfile import time from contextlib import contextmanager TMP_CONTENT_TABLE = 'tmp_content' psycopg2.extras.register_uuid() def stored_procedure(stored_proc): """decorator to execute remote stored procedure, specified as argument Generally, the body of the decorated function should be empty. If it is not, the stored procedure will be executed first; the function body then. """ def wrap(meth): @functools.wraps(meth) def _meth(self, *args, **kwargs): cur = kwargs.get('cur', None) self._cursor(cur).execute('SELECT %s()' % stored_proc) meth(self, *args, **kwargs) return _meth return wrap def jsonize(value): """Convert a value to a psycopg2 JSON object if necessary""" if isinstance(value, dict): return psycopg2.extras.Json(value) return value def entry_to_bytes(entry): """Convert an entry coming from the database to bytes""" if isinstance(entry, memoryview): return entry.tobytes() if isinstance(entry, list): return [entry_to_bytes(value) for value in entry] return entry def line_to_bytes(line): """Convert a line coming from the database to bytes""" if isinstance(line, dict): return {k: entry_to_bytes(v) for k, v in line.items()} return line.__class__(entry_to_bytes(entry) for entry in line) def cursor_to_bytes(cursor): """Yield all the data from a cursor as bytes""" yield from (line_to_bytes(line) for line in cursor) class Db: """Proxy to the SWH DB, with wrappers around stored procedures """ @classmethod def connect(cls, *args, **kwargs): """factory method to create a DB proxy Accepts all arguments of psycopg2.connect; only some specific possibilities are reported below. Args: connstring: libpq2 connection string """ conn = psycopg2.connect(*args, **kwargs) return cls(conn) def _cursor(self, cur_arg): """get a cursor: from cur_arg if given, or a fresh one otherwise meant to avoid boilerplate if/then/else in methods that proxy stored procedures """ if cur_arg is not None: return cur_arg # elif self.cur is not None: # return self.cur else: return self.conn.cursor() def __init__(self, conn): """create a DB proxy Args: conn: psycopg2 connection to the SWH DB """ self.conn = conn @contextmanager def transaction(self): """context manager to execute within a DB transaction Yields: a psycopg2 cursor """ with self.conn.cursor() as cur: try: yield cur self.conn.commit() except: if not self.conn.closed: self.conn.rollback() raise def mktemp(self, tblname, cur=None): self._cursor(cur).execute('SELECT swh_mktemp(%s)', (tblname,)) def mktemp_dir_entry(self, entry_type, cur=None): self._cursor(cur).execute('SELECT swh_mktemp_dir_entry(%s)', (('directory_entry_%s' % entry_type),)) @stored_procedure('swh_mktemp_revision') def mktemp_revision(self, cur=None): pass @stored_procedure('swh_mktemp_release') def mktemp_release(self, cur=None): pass @stored_procedure('swh_mktemp_occurrence_history') def mktemp_occurrence_history(self, cur=None): pass @stored_procedure('swh_mktemp_entity_lister') def mktemp_entity_lister(self, cur=None): pass @stored_procedure('swh_mktemp_entity_history') def mktemp_entity_history(self, cur=None): pass @stored_procedure('swh_mktemp_bytea') def mktemp_bytea(self, cur=None): pass def copy_to(self, items, tblname, columns, cur=None, item_cb=None): def escape(data): if data is None: return '' if isinstance(data, bytes): return '\\x%s' % binascii.hexlify(data).decode('ascii') elif isinstance(data, str): return '"%s"' % data.replace('"', '""') elif isinstance(data, datetime.datetime): # We escape twice to make sure the string generated by # isoformat gets escaped return escape(data.isoformat()) elif isinstance(data, dict): return escape(json.dumps(data)) elif isinstance(data, list): return escape("{%s}" % ','.join(escape(d) for d in data)) elif isinstance(data, psycopg2.extras.Range): # We escape twice here too, so that we make sure # everything gets passed to copy properly return escape( '%s%s,%s%s' % ( '[' if data.lower_inc else '(', '-infinity' if data.lower_inf else escape(data.lower), 'infinity' if data.upper_inf else escape(data.upper), ']' if data.upper_inc else ')', ) ) else: # We don't escape here to make sure we pass literals properly return str(data) with tempfile.TemporaryFile('w+') as f: for d in items: if item_cb is not None: item_cb(d) line = [escape(d.get(k)) for k in columns] f.write(','.join(line)) f.write('\n') f.seek(0) self._cursor(cur).copy_expert('COPY %s (%s) FROM STDIN CSV' % ( tblname, ', '.join(columns)), f) @stored_procedure('swh_content_add') def content_add_from_temp(self, cur=None): pass @stored_procedure('swh_directory_add') def directory_add_from_temp(self, cur=None): pass @stored_procedure('swh_skipped_content_add') def skipped_content_add_from_temp(self, cur=None): pass @stored_procedure('swh_revision_add') def revision_add_from_temp(self, cur=None): pass @stored_procedure('swh_release_add') def release_add_from_temp(self, cur=None): pass @stored_procedure('swh_occurrence_history_add') def occurrence_history_add_from_temp(self, cur=None): pass @stored_procedure('swh_entity_history_add') def entity_history_add_from_temp(self, cur=None): pass def store_tmp_bytea(self, ids, cur=None): """Store the given identifiers in a new tmp_bytea table""" cur = self._cursor(cur) self.mktemp_bytea(cur) self.copy_to(({'id': elem} for elem in ids), 'tmp_bytea', ['id'], cur) def content_missing_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute("""SELECT sha1, sha1_git, sha256 FROM swh_content_missing()""") yield from cursor_to_bytes(cur) def content_missing_per_sha1_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute("""SELECT * FROM swh_content_missing_per_sha1()""") yield from cursor_to_bytes(cur) def skipped_content_missing_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute("""SELECT sha1, sha1_git, sha256 FROM swh_skipped_content_missing()""") yield from cursor_to_bytes(cur) def occurrence_get(self, origin_id, cur=None): """Retrieve latest occurrence's information by origin_id. """ cur = self._cursor(cur) cur.execute("""SELECT origin, branch, target, target_type, (select max(date) from origin_visit where origin=%s) as date FROM occurrence WHERE origin=%s """, (origin_id, origin_id)) yield from cursor_to_bytes(cur) def content_find(self, sha1=None, sha1_git=None, sha256=None, cur=None): """Find the content optionally on a combination of the following checksums sha1, sha1_git or sha256. Args: sha1: sha1 content git_sha1: the sha1 computed `a la git` sha1 of the content sha256: sha256 content Returns: The triplet (sha1, sha1_git, sha256) if found or None. """ cur = self._cursor(cur) cur.execute("""SELECT sha1, sha1_git, sha256, length, ctime, status FROM swh_content_find(%s, %s, %s) LIMIT 1""", (sha1, sha1_git, sha256)) content = line_to_bytes(cur.fetchone()) if set(content) == {None}: return None else: return content def content_find_occurrence(self, sha1, cur=None): """Find one content's occurrence. Args: sha1: sha1 content cur: cursor to use Returns: One occurrence for that particular sha1 """ cur = self._cursor(cur) cur.execute("""SELECT origin_type, origin_url, branch, target, target_type, path FROM swh_content_find_occurrence(%s) LIMIT 1""", (sha1, )) return line_to_bytes(cur.fetchone()) def directory_get_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute('''SELECT id, file_entries, dir_entries, rev_entries FROM swh_directory_get()''') yield from cursor_to_bytes(cur) def directory_missing_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute('SELECT * FROM swh_directory_missing()') yield from cursor_to_bytes(cur) def directory_walk_one(self, directory, cur=None): cur = self._cursor(cur) cur.execute('SELECT * FROM swh_directory_walk_one(%s)', (directory,)) yield from cursor_to_bytes(cur) def directory_walk(self, directory, cur=None): cur = self._cursor(cur) cur.execute('SELECT * FROM swh_directory_walk(%s)', (directory,)) yield from cursor_to_bytes(cur) def revision_missing_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute('SELECT id FROM swh_revision_missing() as r(id)') yield from cursor_to_bytes(cur) revision_add_cols = [ 'id', 'date', 'date_offset', 'date_neg_utc_offset', 'committer_date', 'committer_date_offset', 'committer_date_neg_utc_offset', 'type', 'directory', 'message', 'author_fullname', 'author_name', 'author_email', 'committer_fullname', 'committer_name', 'committer_email', 'metadata', 'synthetic', ] revision_get_cols = revision_add_cols + [ 'author_id', 'committer_id', 'parents'] origin_visit_get_cols = [ 'origin', 'visit', 'date' ] def origin_visit_get(self, origin_id, cur=None): """Retrieve occurrence's history information by origin_id. Args: origin_id: The occurrence's origin Yields: The occurrence's history visits """ cur = self._cursor(cur) cur.execute( 'SELECT origin, visit, date FROM origin_visit where origin=%s', (origin_id, )) yield from cursor_to_bytes(cur) def revision_get_from_temp(self, cur=None): cur = self._cursor(cur) query = 'SELECT %s FROM swh_revision_get()' % ( ', '.join(self.revision_get_cols)) cur.execute(query) yield from cursor_to_bytes(cur) def revision_log(self, root_revisions, limit=None, cur=None): cur = self._cursor(cur) query = """SELECT %s FROM swh_revision_log(%%s, %%s) """ % ', '.join(self.revision_get_cols) cur.execute(query, (root_revisions, limit)) yield from cursor_to_bytes(cur) revision_shortlog_cols = ['id', 'parents'] def revision_shortlog(self, root_revisions, limit=None, cur=None): cur = self._cursor(cur) query = """SELECT %s FROM swh_revision_list(%%s, %%s) """ % ', '.join(self.revision_shortlog_cols) cur.execute(query, (root_revisions, limit)) yield from cursor_to_bytes(cur) def release_missing_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute('SELECT id FROM swh_release_missing() as r(id)') yield from cursor_to_bytes(cur) object_find_by_sha1_git_cols = ['sha1_git', 'type', 'id', 'object_id'] def object_find_by_sha1_git(self, ids, cur=None): cur = self._cursor(cur) self.store_tmp_bytea(ids, cur) query = 'select %s from swh_object_find_by_sha1_git()' % ( ', '.join(self.object_find_by_sha1_git_cols) ) cur.execute(query) yield from cursor_to_bytes(cur) def stat_counters(self, cur=None): cur = self._cursor(cur) cur.execute('SELECT * FROM swh_stat_counters()') yield from cur fetch_history_cols = ['origin', 'date', 'status', 'result', 'stdout', 'stderr', 'duration'] def create_fetch_history(self, fetch_history, cur=None): """Create a fetch_history entry with the data in fetch_history""" cur = self._cursor(cur) query = '''INSERT INTO fetch_history (%s) VALUES (%s) RETURNING id''' % ( ','.join(self.fetch_history_cols), ','.join(['%s'] * len(self.fetch_history_cols)) ) cur.execute(query, [fetch_history.get(col) for col in self.fetch_history_cols]) return cur.fetchone()[0] def get_fetch_history(self, fetch_history_id, cur=None): """Get a fetch_history entry with the given id""" cur = self._cursor(cur) query = '''SELECT %s FROM fetch_history WHERE id=%%s''' % ( ', '.join(self.fetch_history_cols), ) cur.execute(query, (fetch_history_id,)) data = cur.fetchone() if not data: return None ret = {'id': fetch_history_id} for i, col in enumerate(self.fetch_history_cols): ret[col] = data[i] return ret def update_fetch_history(self, fetch_history, cur=None): """Update the fetch_history entry from the data in fetch_history""" cur = self._cursor(cur) query = '''UPDATE fetch_history SET %s WHERE id=%%s''' % ( ','.join('%s=%%s' % col for col in self.fetch_history_cols) ) cur.execute(query, [jsonize(fetch_history.get(col)) for col in self.fetch_history_cols + ['id']]) base_entity_cols = ['uuid', 'parent', 'name', 'type', 'description', 'homepage', 'active', 'generated', 'lister_metadata', 'metadata'] entity_cols = base_entity_cols + ['last_seen', 'last_id'] entity_history_cols = base_entity_cols + ['id', 'validity'] def origin_add(self, type, url, cur=None): """Insert a new origin and return the new identifier.""" insert = """INSERT INTO origin (type, url) values (%s, %s) RETURNING id""" cur.execute(insert, (type, url)) return cur.fetchone()[0] def origin_get_with(self, type, url, cur=None): """Retrieve the origin id from its type and url if found.""" cur = self._cursor(cur) query = """SELECT id, type, url, lister, project FROM origin WHERE type=%s AND url=%s""" cur.execute(query, (type, url)) data = cur.fetchone() if data: return line_to_bytes(data) return None def origin_get(self, id, cur=None): """Retrieve the origin per its identifier. """ cur = self._cursor(cur) query = "SELECT id, type, url, lister, project FROM origin WHERE id=%s" cur.execute(query, (id,)) data = cur.fetchone() if data: return line_to_bytes(data) return None person_cols = ['fullname', 'name', 'email'] person_get_cols = person_cols + ['id'] def person_add(self, person, cur=None): """Add a person identified by its name and email. Returns: The new person's id """ cur = self._cursor(cur) query_new_person = '''\ INSERT INTO person(%s) VALUES (%s) RETURNING id''' % ( ', '.join(self.person_cols), ', '.join('%s' for i in range(len(self.person_cols))) ) cur.execute(query_new_person, [person[col] for col in self.person_cols]) return cur.fetchone()[0] def person_get(self, ids, cur=None): """Retrieve the persons identified by the list of ids. """ cur = self._cursor(cur) query = """SELECT %s FROM person WHERE id IN %%s""" % ', '.join(self.person_get_cols) cur.execute(query, (tuple(ids),)) yield from cursor_to_bytes(cur) release_add_cols = [ 'id', 'target', 'target_type', 'date', 'date_offset', 'date_neg_utc_offset', 'name', 'comment', 'synthetic', 'author_fullname', 'author_name', 'author_email', ] release_get_cols = release_add_cols + ['author_id'] def release_get_from_temp(self, cur=None): cur = self._cursor(cur) query = ''' SELECT %s FROM swh_release_get() ''' % ', '.join(self.release_get_cols) cur.execute(query) yield from cursor_to_bytes(cur) def release_get_by(self, origin_id, limit=None, cur=None): """Retrieve a release by occurrence criterion (only origin right now) Args: - origin_id: The origin to look for. """ cur = self._cursor(cur) query = """ SELECT %s FROM swh_release_get_by(%%s) LIMIT %%s """ % ', '.join(self.release_get_cols) cur.execute(query, (origin_id, limit)) yield from cursor_to_bytes(cur) def revision_get_by(self, origin_id, branch_name, datetime, limit=None, cur=None): """Retrieve a revision by occurrence criterion. Args: - origin_id: The origin to look for - branch_name: the branch name to look for - datetime: the lower bound of timerange to look for. - limit: limit number of results to return The upper bound being now. """ cur = self._cursor(cur) if branch_name and isinstance(branch_name, str): branch_name = branch_name.encode('utf-8') query = ''' SELECT %s FROM swh_revision_get_by(%%s, %%s, %%s) LIMIT %%s ''' % ', '.join(self.revision_get_cols) cur.execute(query, (origin_id, branch_name, datetime, limit)) yield from cursor_to_bytes(cur) def directory_entry_get_by_path(self, directory, paths, cur=None): """Retrieve a directory entry by path. """ cur = self._cursor(cur) cur.execute("""SELECT dir_id, type, target, name, perms, status, sha1, sha1_git, sha256 FROM swh_find_directory_entry_by_path(%s, %s)""", (directory, paths)) data = cur.fetchone() if set(data) == {None}: return None return line_to_bytes(data) def entity_get(self, uuid, cur=None): """Retrieve the entity and its parent hierarchy chain per uuid. """ cur = self._cursor(cur) cur.execute("""SELECT %s FROM swh_entity_get(%%s)""" % ( ', '.join(self.entity_cols)), (uuid, )) yield from cursor_to_bytes(cur) def entity_get_one(self, uuid, cur=None): """Retrieve a single entity given its uuid. """ cur = self._cursor(cur) cur.execute("""SELECT %s FROM entity WHERE uuid = %%s""" % ( ', '.join(self.entity_cols)), (uuid, )) data = cur.fetchone() if not data: return None return line_to_bytes(data) def archive_ls(self, cur=None): """ Get all the archives registered on the server. Yields: a tuple (server_id, server_url) for each archive server. """ cur = self._cursor(cur) cur.execute("""SELECT id, url FROM archive """) yield from cursor_to_bytes(cur) def content_archive_get(self, content=None, cur=None): """ Get the archival status of a content in a specific server. Retrieve from the database the archival status of the given content in the given archive server. Args: content: the sha1 of the content. May be None for all contents. Yields: A tuple (content_id, copies_json). """ query = """SELECT content_id, copies FROM content_archive """ if content is not None: query += "WHERE content_id='%s'" % content else: query += 'ORDER BY content_id' cur = self._cursor(cur) cur.execute(query) yield from cursor_to_bytes(cur) + def content_archive_get_copies(self, previous_content=None, limit=1000, + cur=None): + """Get the list of copies for `limit` contents starting after + `previous_content`. + + Args: + previous_content: sha1 of the last content retrieved. May be None + to start at the beginning. + limit: number of contents to retrieve. Can be None to retrieve all + objects (will be slow). + + Yields: + A tuple (content_id, present_copies, ongoing_copies), where + ongoing_copies is a dict mapping copy to mtime. + + """ + + query = """SELECT content_id, + array( + SELECT key + FROM jsonb_each(copies) + WHERE value->>'status' = 'present' + ORDER BY key + ) AS present, + array( + SELECT key + FROM jsonb_each(copies) + WHERE value->>'status' = 'ongoing' + ORDER BY key + ) AS ongoing, + array( + SELECT value->'mtime' + FROM jsonb_each(copies) + WHERE value->>'status' = 'ongoing' + ORDER BY key + ) AS ongoing_mtime + FROM content_archive + WHERE content_id > %s + ORDER BY content_id + LIMIT %s + """ + + if previous_content is None: + previous_content = b'' + + cur = self._cursor(cur) + cur.execute(query, (previous_content, limit)) + for content_id, present, ongoing, mtimes in cursor_to_bytes(cur): + yield (content_id, present, dict(zip(ongoing, mtimes))) + def content_archive_update(self, content_id, archive_id, new_status=None, cur=None): """ Update the status of an archive content and set its mtime to Change the mtime of an archived content for the given archive and set it's mtime to the current time. Args: content_id (str): content sha1 archive_id (str): name of the archive new_status (str): one of 'missing', 'present' or 'ongoing'. this status will replace the previous one. If not given, the function only change the mtime of the content for the given archive. """ if new_status is not None: query = """UPDATE content_archive SET copies=jsonb_set( copies, '{%s}', '{"status":"%s", "mtime":%d}' ) WHERE content_id='%s' """ % (archive_id, new_status, int(time.time()), content_id) else: query = """ UPDATE content_archive SET copies=jsonb_set(copies, '{%s,mtime}', '%d') WHERE content_id='%s' """ % (archive_id, int(time.time())) cur = self._cursor(cur) cur.execute(query)